Skip to content

Draft v0.1

Summary

Lifecycle Stage 6
This article is part of the AI Interaction Lifecycle framework
Express Intent → Display Results → Refine Output → Take Action → Execute & Monitor → Summary

Overview

As AI interactions become longer and more complex, users often need a concise understanding of what occurred during a conversation, workflow, or execution process.

Summary focuses on the patterns that help users quickly understand outcomes, decisions, completed actions, supporting evidence, and recommended next steps without requiring them to review an entire interaction history.

Whether summarizing a conversation, a research task, a workflow, or an autonomous agent execution, effective summaries reduce cognitive load, improve transparency, and help users maintain continuity between interactions.

The goal of this stage is to help users understand what was accomplished, verify important information, and determine what should happen next.

Patterns

Outcome Summary

Outcome Summaries provide a concise overview of what was accomplished during an interaction.

An outcome summary may include:

  • Goals achieved
  • Key findings
  • Decisions made
  • Content created
  • Actions completed
  • Outstanding work

Outcome summaries help users quickly understand results and communicate progress to others.

Examples include:

  • AI-generated meeting recaps
  • Research summaries
  • Agent execution reports
  • Completed task summaries

Explainability Layers

Explainability Layers provide additional context that helps users understand how conclusions, recommendations, or actions were generated.

Rather than exposing every detail by default, explanations can be progressively disclosed based on user needs.

These layers may reveal:

  • Sources consulted
  • Supporting evidence
  • Reasoning summaries
  • Actions performed
  • Decision points
  • Confidence indicators

Explainability helps users understand not only what happened, but why it happened.

Action Plan Recap

Action Plan Recaps provide a summary of the actions an AI planned, executed, or recommended throughout a workflow.

Depending on the experience, summaries may include:

  • Planned actions
  • Completed actions
  • Skipped actions
  • Failed actions
  • User approvals
  • Outstanding tasks
  • Recommended next steps

Action plan recaps help users understand how outcomes were achieved and identify any remaining work.

Variations

Step Lists

Sequential summaries of actions that were planned or completed.

Execution Previews

Structured plans tied to code or automation, with explicit approval gates.

Content Outlines

Document or slide scaffolds that may or may not require confirmation.

Adaptive Plans

Plans that evolve mid-process, sometimes with repeated confirmations.

Citations

Citations connect summarized information back to supporting evidence and source material.

They help users:

  • Verify information
  • Review supporting context
  • Understand where information originated
  • Investigate important details when necessary

The primary goal of citations is to support verification and maintain confidence in summarized outputs.

Variations

Inline Highlights

Best suited for attached content such as PDFs and documents.

Example:
Adobe Acrobat highlights source passages directly alongside generated summaries.
Direct Quotations

Best suited for transcripts, meeting notes, and long-form content.

Example:
Granola associates key takeaways with transcript excerpts.
Multi-Source References

Best suited for research and aggregation experiences.

Example:
Perplexity includes citations from multiple sources within its summaries.

Best suited for simple verification scenarios.

Example:
Copy.ai provides source links for generated content.

Principles Applied

This stage primarily aligns with the following AI UX principles:

  • Transparency – Users should be able to understand what occurred, what actions were taken, and how outcomes were achieved.
  • Trust & Confidence – Users should be able to evaluate summarized information and act upon it appropriately.
  • Accountability – Actions, decisions, approvals, and outcomes should be understandable, reviewable, and traceable.

Next Steps

Next Steps help users understand what actions may be required after a workflow, conversation, or execution process has completed.

Recommendations may include:

  • Follow-up actions
  • Outstanding approvals
  • Additional research opportunities
  • Workflow continuation points
  • Suggested refinements

By providing clear recommendations, AI systems help users efficiently transition from completed work to future tasks.

End of the AI Interaction Lifecycle

The Summary stage concludes the AI Interaction Lifecycle.

By this point, users have:

  • Expressed intent
  • Evaluated results
  • Refined outcomes
  • Planned actions
  • Monitored execution
  • Reviewed outcomes

Together, these six stages provide a framework for designing AI experiences that support both informational interactions and agentic workflows while maintaining transparency, trust, control, and accountability throughout the user journey.